DESCRIPTION: (Applicant's Abstract) Injection drug users (IDU) constitute an increasing proportion of new cases of AIDS and HIV infection. Yet little research has been conducted on the utilization of health services by active injection drug users. In the proposed project, 500 IDUs in Providence, Rhode Island will be interviewed regarding their health service use in the past six months and these self-reports will then be validated by abstraction of inpatient, outpatient and emergency room records at three hospitals in Providence. The study sample will include 250 active IDUs and 250 former injection drug users who have been in methadone maintenance for at least six months prior to enrollment. This retrospective cohort design takes advantage of the geographic compactness of Rhode Island, a single Needle Exchange program for recruitment of active drug users, limited out-migration of IDUs, the small number of health care facilities in our state and investigators at each of these facilities who have previously worked collaboratively on epidemiologic studies. Specific Aim 1 is to provide estimates of health service utilization (emergency room, inpatient, outpatient) among different groups of IDUs--with or without HIV infection, in drug treatment and active users. We will test the following hypotheses: IDUs in drug treatment will have lower utilization of expensive health services (ER, hospitalization) than active IDUs; cocaine users will have higher rates of emergency room use and hospitalization than IDUs who do not use cocaine both in the active IDU cohort and the methadone cohort due to more frequent drug use and injection. Specific Aim 2 is to provide estimates of total costs for IDUs in and out of treatment and test the hypothesis that the total cost for health care for IDUs will be lower among those receiving methadone maintenance. The medico-economic consequences of injection drug use will be important for Medicaid, and for Health Maintenance Organizations that are assuming Medicaid contracts and that currently have little data on which to base their benefit and program design decisions for this population